I made this lil bluetooth speaker! I took it to Burning Amp 2022 and a few people were entertained by it, so I figured I'd put up some info about it. Also, it's fairly clearly not fully optimized, so as I continue to work on it ... slowly ... I'll post up results here.
First! Every thread is better with pictures, so here it is.
This is what I built based on driver suggestion from Planet10 back in this thread https://www.diyaudio.com/community/threads/driver-choice-for-bt-speaker.385079/#post-6995827 -- if admins wish to combine these threads that's fine. I thought about just continuing with that thread but felt the thread name (Driver choice for BT speaker) was misleading since I've now bought the drivers and am no longer seeking input there.
So what's going on here? This is a pair of MarkAudio CHN-50, in 2 liters per side, each tuned to target 80 Hz. I'm using a Dayton KABD-250 to power them. It's cool b/c it's got bluetooth built in, nice enough little amp, and it's got DSP and uses Sigma Studio which I was excited to try out. It's also got line-in and line-out. There's a battery board that matches it, and you need to get a programmer board to work the DSP angle.
As we played with it at BA, it's got 4th order HP at 80 Hz, and some compression (dual band! yay for Sigma Studio's flexibility) and some EQ. Honestly I don't feel like I've gotten a really complete understanding of the FR yet (I've made a few measurements) so the EQ probably needs honing.
Further, BassBox said that 2 liters would take a port 5" wide and 1/4" tall and 7 1/8 inches long (with a 90 deg turn internal) to tune to 80 Hz. Impedance sweep made it look like I had hit the target, but the guy who was measuring stuff at BA took a nearfield of the driver and said he saw the port null down at 60. So I'll have to verify that and change the porting if necessary, as matching the HP and the tuning frequency to save excursion is a must. Why the discrepency? Maybe the substantial aspect ratio of the port makes it function differently at low level (impedance sweep by DATS) vs higher level (measurement, music playback)? Needs investigation.
Further Mr. Measurement was getting reasonable FR, but REW was giving him odd looking impulse. AND, he came to the conclusion that there was something funny going on, he said he could feel the off cone being driven when feeding a single channel into the measurement (connection over BT), so a few people thought something funny going on with some processing in the BT setup (as distinct from the DSP chip), so I'll have to check that out also.
Finally, if I can get everything else together, the plan is to use flexible tuning options in Sigma Studio to make a sub out with the line-out, that can switch to a higher HP optionally and feed everything else to an external sub.
So that's what's up! I got a job and a family and a house, so probably gonna take some time to go through all this, but I wanted to start a thread while the folks I met at BA maybe still had an idea of the device.
First! Every thread is better with pictures, so here it is.
This is what I built based on driver suggestion from Planet10 back in this thread https://www.diyaudio.com/community/threads/driver-choice-for-bt-speaker.385079/#post-6995827 -- if admins wish to combine these threads that's fine. I thought about just continuing with that thread but felt the thread name (Driver choice for BT speaker) was misleading since I've now bought the drivers and am no longer seeking input there.
So what's going on here? This is a pair of MarkAudio CHN-50, in 2 liters per side, each tuned to target 80 Hz. I'm using a Dayton KABD-250 to power them. It's cool b/c it's got bluetooth built in, nice enough little amp, and it's got DSP and uses Sigma Studio which I was excited to try out. It's also got line-in and line-out. There's a battery board that matches it, and you need to get a programmer board to work the DSP angle.
As we played with it at BA, it's got 4th order HP at 80 Hz, and some compression (dual band! yay for Sigma Studio's flexibility) and some EQ. Honestly I don't feel like I've gotten a really complete understanding of the FR yet (I've made a few measurements) so the EQ probably needs honing.
Further, BassBox said that 2 liters would take a port 5" wide and 1/4" tall and 7 1/8 inches long (with a 90 deg turn internal) to tune to 80 Hz. Impedance sweep made it look like I had hit the target, but the guy who was measuring stuff at BA took a nearfield of the driver and said he saw the port null down at 60. So I'll have to verify that and change the porting if necessary, as matching the HP and the tuning frequency to save excursion is a must. Why the discrepency? Maybe the substantial aspect ratio of the port makes it function differently at low level (impedance sweep by DATS) vs higher level (measurement, music playback)? Needs investigation.
Further Mr. Measurement was getting reasonable FR, but REW was giving him odd looking impulse. AND, he came to the conclusion that there was something funny going on, he said he could feel the off cone being driven when feeding a single channel into the measurement (connection over BT), so a few people thought something funny going on with some processing in the BT setup (as distinct from the DSP chip), so I'll have to check that out also.
Finally, if I can get everything else together, the plan is to use flexible tuning options in Sigma Studio to make a sub out with the line-out, that can switch to a higher HP optionally and feed everything else to an external sub.
So that's what's up! I got a job and a family and a house, so probably gonna take some time to go through all this, but I wanted to start a thread while the folks I met at BA maybe still had an idea of the device.
Ok, so it's been a while!
It was troubling to me, the idea that the Dayton amp KABD-250 might have some hocus pocus going on in it that I was not in control of. This was the suggestion of some folks coming out of BA'22. So I have measured it electrically. My conclusion - nothing untoward is going on.
If you look at the picture, you'll see the two ports come out right next to each other. So I think that in the case where one channel is driven and the other not, the two ports couple together. This allows, particularly at port resonance, one driver to drive the other, and this is what the guest measurement expert at BA felt.
So this an electrical measurement, graph is dBV vs Frequency. The top traces are the output of the amp, with EQ in place as I currently have it. I tried driving only one channel, and measuring the output on both sides. The driven side reproduced the top curve, and the un-driven side gave the trace which peaks around -50 dBV @ 80 Hz. If you were looking for some sort of digital widening / spatialization, I would sort of expect to see it higher up in frequency. Why 80 Hz instead of 60 (where tuning null had been spotted)? Maybe because of the EQ shape? Or I seem to recall the impedance curve looking like an 80 Hz tuning while the null looked lower, maybe something with that? Either way, once I disconnected the driver and re-measured (on the amp side), the 80 Hz energy disappeared, and I got the pinkish curve. A bit of noise floor bulge probably representing actual electrical crosstalk from the EQ. But with the un-driven speaker disconnected, a measurement of the terminals of the driver (into no load) made the lavender curve, peaking at -20 dBV @ 75 Hz. So that's clearly picking up energy mechanically from the driven channel.
What can I say on these trace colors, it's an excel to LibreOffice palette thing. It looks like CGA color palette from 1987.
So my concerns are hereby assuaged that the KABD-250 is performing some spatialization or other crazy processing behind my back.
For my next trick, I will investigate what is going on with the tuning frequency, and perhaps saw the face off the device in order to re-implement a port that meets the target frequency. That will make for less cone excursion above the high pass, and less LF EQ boost making for a easier time for the amp and limiters. I've gotten a 3D printer since the start of this project, so I can start printing tube ports instead of designing slot ports into the box. I'm excited to play with that.
Until next time!
- A
It was troubling to me, the idea that the Dayton amp KABD-250 might have some hocus pocus going on in it that I was not in control of. This was the suggestion of some folks coming out of BA'22. So I have measured it electrically. My conclusion - nothing untoward is going on.
If you look at the picture, you'll see the two ports come out right next to each other. So I think that in the case where one channel is driven and the other not, the two ports couple together. This allows, particularly at port resonance, one driver to drive the other, and this is what the guest measurement expert at BA felt.
So this an electrical measurement, graph is dBV vs Frequency. The top traces are the output of the amp, with EQ in place as I currently have it. I tried driving only one channel, and measuring the output on both sides. The driven side reproduced the top curve, and the un-driven side gave the trace which peaks around -50 dBV @ 80 Hz. If you were looking for some sort of digital widening / spatialization, I would sort of expect to see it higher up in frequency. Why 80 Hz instead of 60 (where tuning null had been spotted)? Maybe because of the EQ shape? Or I seem to recall the impedance curve looking like an 80 Hz tuning while the null looked lower, maybe something with that? Either way, once I disconnected the driver and re-measured (on the amp side), the 80 Hz energy disappeared, and I got the pinkish curve. A bit of noise floor bulge probably representing actual electrical crosstalk from the EQ. But with the un-driven speaker disconnected, a measurement of the terminals of the driver (into no load) made the lavender curve, peaking at -20 dBV @ 75 Hz. So that's clearly picking up energy mechanically from the driven channel.
What can I say on these trace colors, it's an excel to LibreOffice palette thing. It looks like CGA color palette from 1987.
So my concerns are hereby assuaged that the KABD-250 is performing some spatialization or other crazy processing behind my back.
For my next trick, I will investigate what is going on with the tuning frequency, and perhaps saw the face off the device in order to re-implement a port that meets the target frequency. That will make for less cone excursion above the high pass, and less LF EQ boost making for a easier time for the amp and limiters. I've gotten a 3D printer since the start of this project, so I can start printing tube ports instead of designing slot ports into the box. I'm excited to play with that.
Until next time!
- A
Update time!
So Scott H's measurements (back at BA '22) had indicated a null in speaker performance down at 60 Hz, but I had intended to tune it at 80 Hz. And it was an awfully narrow slot port. It was just the easiest way to build it, but not really optimal. And I've gotten this 3D printer and started using FreeCad. So I figured I'd do up this port situation.
Here's the new hotness:
Aww yeah!
So there's this Harman paper, "Maximizing Performance from Loudspeaker Ports". Some google gives me a DIYAudio link - https://www.diyaudio.com/community/attachments/harman-port-study-pdf.893225/
And that talks about some geometry and has measurements, and I don't know if it's the BEST best? But it's the best I know of at the moment. And the shape of it was something I figured I could make in CAD. So I did, and then printed it on my 3D printer and here it is.
Nice. I made them in two parts that I can glue together to keep printing supports down, and quality up. Here's a nearfield of the transducer, and the port..
SPL levels aren't calibrated. Hitting that 80 Hz!
I spent some time measuring on it, and came up with this EQ...
Here's a ground plane - no EQ, then with EQ...
I had made up EQ for it before. After the port change I figured I'd need to alter the bass boost (still needs some b/c the chamber is undersized), and I went over the rest of it. I might back off on that 90 Hz, I'll listen for a while and see if I think it's too much. I kind of feel like I need to measure stuff in many different ways to see what changes and what stays the same. Maybe if I had a better space for acoustic measurements I'd be able to give it a single go and trust what I had. But doing it around the house it's all compromised, so it's just a question of finding a technique that leaves the artifacts out of the way. This one was a groundplane, in the driveway, 42 inches from the cone (3x the baffle width)
I'm kinda bummin about that hole at 2.2 kHz. I didn't want to throw energy into it, so I didn't try to EQ it up. But that's at 2.2 kHz, so what's the quarter wavelength? 6 inches. =( That's the front to back wavelength of my acoustic chamber. So I'll have to consider. I don't want to just stuff it to death, I want a good pathway for that fancy new port to work. Hmmmm....
Anyway it sounds pretty good at this point. Not much image with the drivers so close together, but if you can sit close it makes a little sweet spot that sounds hi-fi-ish. Wandering around the room it sounds good for a BT speaker. It plays quite loudly for what it is in a domestic environment. I think the battery pack only gives it a dozen volts, so the amp can't really get out of hand? But maybe I should measure that, they might have a boost circuit in there or something.
So Scott H's measurements (back at BA '22) had indicated a null in speaker performance down at 60 Hz, but I had intended to tune it at 80 Hz. And it was an awfully narrow slot port. It was just the easiest way to build it, but not really optimal. And I've gotten this 3D printer and started using FreeCad. So I figured I'd do up this port situation.
Here's the new hotness:
Aww yeah!
So there's this Harman paper, "Maximizing Performance from Loudspeaker Ports". Some google gives me a DIYAudio link - https://www.diyaudio.com/community/attachments/harman-port-study-pdf.893225/
And that talks about some geometry and has measurements, and I don't know if it's the BEST best? But it's the best I know of at the moment. And the shape of it was something I figured I could make in CAD. So I did, and then printed it on my 3D printer and here it is.
Nice. I made them in two parts that I can glue together to keep printing supports down, and quality up. Here's a nearfield of the transducer, and the port..
SPL levels aren't calibrated. Hitting that 80 Hz!
I spent some time measuring on it, and came up with this EQ...
Here's a ground plane - no EQ, then with EQ...
I had made up EQ for it before. After the port change I figured I'd need to alter the bass boost (still needs some b/c the chamber is undersized), and I went over the rest of it. I might back off on that 90 Hz, I'll listen for a while and see if I think it's too much. I kind of feel like I need to measure stuff in many different ways to see what changes and what stays the same. Maybe if I had a better space for acoustic measurements I'd be able to give it a single go and trust what I had. But doing it around the house it's all compromised, so it's just a question of finding a technique that leaves the artifacts out of the way. This one was a groundplane, in the driveway, 42 inches from the cone (3x the baffle width)
I'm kinda bummin about that hole at 2.2 kHz. I didn't want to throw energy into it, so I didn't try to EQ it up. But that's at 2.2 kHz, so what's the quarter wavelength? 6 inches. =( That's the front to back wavelength of my acoustic chamber. So I'll have to consider. I don't want to just stuff it to death, I want a good pathway for that fancy new port to work. Hmmmm....
Anyway it sounds pretty good at this point. Not much image with the drivers so close together, but if you can sit close it makes a little sweet spot that sounds hi-fi-ish. Wandering around the room it sounds good for a BT speaker. It plays quite loudly for what it is in a domestic environment. I think the battery pack only gives it a dozen volts, so the amp can't really get out of hand? But maybe I should measure that, they might have a boost circuit in there or something.
I would love to using the FR as sealed and upto 120Hz and use a 8" or 6.5" sub driver and a 2.1 amp
It's a nice little driver! I bet it would do well as a mid-tweet in a 2-way. I've set up the output on the KABD as a sub out (have to set a low pass at high order! try to match 4th order HP on the speaker, with 4th order for the vented box), and I have an 8" and a small plate amp that are waiting their turn to become a small sub to go with this BT speaker.